GRAD：基于强化指数移动平均和多阶段滑动窗口技术的自动驾驶车辆传感器实时门控循环异常检测
摘要
本文介绍了GRAD，这是一种用于自动驾驶车辆传感器的实时异常检测方法，集成了统计分析和深度学习，以确保传感器数据的可靠性。该方法结合了自适应异常检测平滑因子和阈值的强化指数移动平均 (REMA)，以及用于捕捉短期和长期模式的多阶段滑动窗口 (MS-SW) 技术。这些特征使用轻量级门控循环单元 (GRU) 模型进行处理，基于偏置类型检测和分类异常，同时通过恢复模块恢复损坏的传感器数据，以确保系统持续运行。GRAD 的轻量级架构由两层 GRU 组成，神经元数量有限，适用于实时应用，同时保持高检测准确率。GRAD 框架在异常检测和分类方面取得了显著的性能。该模型对异常数据和正常数据分别实现了 97.6% 和 99.4% 的总F1分数，凸显其在区分正常和异常传感器数据方面的高准确性。Regarding the anomaly classification, GRAD successfully categorized different anomaly types with high precision, enabling the recovery module to accurately restore damaged sensor data. Relative to analogous studies, GRAD surpasses current models by attaining a balance between elevated detection accuracy and diminished computational expense. These results demonstrate GRAD's potential as a reliable and efficient solution for real-time anomaly detection in autonomous vehicle systems, guaranteeing safe vehicle operation with minimal computational overhead.
引用
@article{arxiv.2510.23327,
title = {GRAD: Real-Time Gated Recurrent Anomaly Detection in Autonomous Vehicle Sensors Using Reinforced EMA and Multi-Stage Sliding Window Techniques},
author = {Mohammad Hossein Jafari Naeimi and Ali Norouzi and Athena Abdi},
journal= {arXiv preprint arXiv:2510.23327},
year = {2025}
}